Alignment and exposure stacks, feeling like a noob

[Bob Mahar]

Hi,

I've been using Hugin for years now.   Perhaps I got amnesia along the way and forgot some critical checkbox I automatically would set.   But recently I've been having some issues with exposure bracketed stacks when making stereographic panoramas.   The fused image will have ghosts where it seems as if the stack members are not alaigned properly and cause ghosting / multiple exposure like effects in the stitched image.

I use a tripod and a wired remote release to take a 5 or 7 exposure bracket for each of the shots.   The images for a given stack do not appear to be misaligned as shot, To cover the world with the 14mm lens, perhaps 25-30 bracketed shots.   I plop them into Hugin.  Set the stack size as appropriate.  Use the cpfind ( rows + stacks ).   Then after verifying everything is connected,  do an alignment for Position to rough it in.  Assuming there is nothing too ridiculous in the control point table distances, go for the Position + Barrel.   Normally I get a very nice result, with limited need to futz or post process.  

However in some recent projects, I see ghosting on the fused output, as if the individual stack members are misaligned.slightly.

Since I'm using a tripod, and fixed aperture and focus, the stack images should already be registered properly, so I should be able to ignore the alignment of individual stack members, and rely on the +/-0 EV images for alignment.   I never noticed this ghosting / multiple images issues in prior projects.   And it does not appear in the ordinary non-fused project - so seems to be an artifact of the multiple exposure layers differing in alaignment and then after being fused you have the ghosting.

So, this is a very log winded way of asking: is there a way to force Hugin to use only control points for one of the exposure bracket members, then apply the distortions and masks necessary to all of the stack members?   Is this a behavior that might have changed in recent history?   I assume this is a very commonly desired behavior, so I must be missing something really obvious in my workflow.

Any suggestions are appreciated.

-- Bob

[Terry Duell]

Hello Bob,

On Wed, 27 May 2015 07:11:12 +1000, Bob Mahar <muhl...@gmail.com> wrote:


>
> So, this is a very log winded way of asking: is there a way to force
> Hugin to use only control points for one of the exposure bracket
> members, then
> apply the distortions and masks necessary to all of the stack members?
> Is this a behavior that might have changed in recent history? I assume
> this is a very commonly desired behavior, so I must be missing something
> really obvious in my workflow.
>

Just to test whether the images in each stack need alignment, I would load
one stack as separate images, and run CPfind and optimise. That should
tell you if each stack needs alignment.
Assuming that your stacks don't need alignment, after loading all your
images and setting the stacks, you can choose an image in each stack,
right mouse to get the context menu, choose stacks -> set stack size, and
set 'link positions of images in each stack'.
You should be able to use the normal CPfind and hugin should only set CPs
in one image of each stack. Hope I have that right!
See if that helps.

Cheers,
--
Regards,
Terry Duell

[bran...@yahoo.com]

What version are you using? Might want to double check that only one image from each stack is being linked to images from other stacks. When I first started I was getting control points from every image in a stack to other stacks and that is a very good way to cause what you are describing.

As far as how hugin works, I am pretty sure if you have the correct settings on the control point detector it should work as I described.(I had enough problems in 2013 version that I made a bat that does all of the control points for me so I do not need to stress about it and have since lost touch with how hugin currently does it.)